Interesting file, but the author has embarked on research that assumes use of sprockets below the minimum tooth count recommended by manufacturers, on the basis that they were stock on a motorcycle.
His tests are limited to 13, 14, 15 and 16 tooth.
Smallest size (i.e., "this is the smallest sprocket you should use for a power train unless there is no alternative"): 17.
Smaller sprockets eat chains and shaft bearings (Honda 750 K is the poster child here), and have huge cyclic pitch changes due to the polygonal effect he mentions.
This is a strong argument for dual row: you have too few teeth on the minor sprocket in your load and ratio choice, and don't have enough load capacity in a single row.
There are also cases where the diameter is simply too big for convenience, and reducing the pitch with dual or triple saves you.
There may be some useful conclusions in his data, but not as much as if sizes chosen by engineers were used instead of what his bike had.
I certainly agree that by whatever method the chain must be wet throughout the run, not only for lube but to carry away heat, or it will simply de-harden and be chewed away like a chocolate bar.
Many industrial apps use an enclosed bath, and the chain industry is very happy to tell you exactly what oil, grease etc. you need for your speed/load/temp.